lable at ScienceDirect

Journal of Human Evolution 112 (2017) 93e104

Contents lists avai

Journal of Human Evolution

journal homepage: www.elsevier .com/locate/ jhevol

Hominin track assemblages from Okote Member deposits near Ileret,Kenya, and their implications for understanding fossil homininpaleobiology at 1.5 Ma

Kevin G. Hatala a, b, *, Neil T. Roach c, d, Kelly R. Ostrofsky b, Roshna E. Wunderlich e,Heather L. Dingwall c, Brian A. Villmoare f, David J. Green g, David R. Braun b,John W.K. Harris h, Anna K. Behrensmeyer i, Brian G. Richmond d, j

a Department of Biology, Chatham University, Pittsburgh, PA 15232, USAb Center for the Advanced Study of Human Paleobiology, Department of Anthropology, The George Washington University, Washington, DC 20052, USAc Department of Human Evolutionary Biology, Harvard University, Cambridge, MA 02138, USAd Division of Anthropology, American Museum of Natural History, New York, NY 10024, USAe Department of Biology, James Madison University, Harrisonburg, VA 22807, USAf Department of Anthropology, University of Nevada Las Vegas, Las Vegas, NV 89154, USAg Department of Anatomy, Midwestern University, Downers Grove, IL 60515, USAh Department of Anthropology, Rutgers University, New Brunswick, NJ 08901, USAi Department of Paleobiology, National Museum of Natural History, Smithsonian Institution, Washington, DC 20013, USAj Humboldt Foundation Fellow at Department of Human Evolution, Max Planck Institute for Evolutionary Anthropology, Leipzig D-04103, Germany

a r t i c l e i n f o

Article history:Received 1 September 2016Accepted 7 August 2017Available online 13 September 2017

Keywords:IchnologyTrace fossilsFootprintsHomo erectusKoobi ForaPleistocene

* Corresponding author.E-mail address: kevin.g.hatala@gmail.com (K.G. Ha

http://dx.doi.org/10.1016/j.jhevol.2017.08.0130047-2484/© 2017 Elsevier Ltd. All rights reserved.

a b s t r a c t

Tracks can provide unique, direct records of behaviors of fossil organisms moving across their landscapesmillions of years ago. While track discoveries have been rare in the human fossil record, over the lastdecade our team has uncovered multiple sediment surfaces within the Okote Member of the Koobi ForaFormation near Ileret, Kenya that contain large assemblages of ~1.5 Ma fossil hominin tracks. Here, weprovide detailed information on the context and nature of each of these discoveries, and we outline thespecific data that are preserved on the Ileret hominin track surfaces. We analyze previously unpublisheddata to refine and expand upon earlier hypotheses regarding implications for hominin anatomy andsocial behavior. While each of the track surfaces discovered at Ileret preserves a different amount of datathat must be handled in particular ways, general patterns are evident. Overall, the analyses presentedhere support earlier interpretations of the ~1.5 Ma Ileret track assemblages, providing further evidence oflarge, human-like body sizes and possibly evidence of a group composition that could support theemergence of certain human-like patterns of social behavior. These data, used in concert with otherforms of paleontological and archaeological evidence that are deposited on different temporal scales,offer unique windows through which we can broaden our understanding of the paleobiology of homi-nins living in East Africa at ~1.5 Ma.

© 2017 Elsevier Ltd. All rights reserved.

1. Introduction

Across the field of paleontology, ichnological studies havefigured prominently in the developments of major evolutionaryhypotheses. Tracks (often referred to as ‘footprints’ in

tala).

paleoanthropological literature) and trackways (sequences of twoor more consecutive tracks) have been used to address a wide arrayof questions related to the biology of fossil organisms includingquestions related to social behavior (e.g., Ostrom,1972; Lockley andMeyer, 1994, 2006; Cotton et al., 1998; Matsukawa et al., 2001;Lingham-Soliar et al., 2003; Bibi et al., 2012), paleoecology (e.g.,Lockley et al., 2007; Dentzien-Dias et al., 2008; Smith et al., 2009;Kukihara and Lockley, 2012), body size (e.g., Lockley, 1994;Henderson, 2003), and locomotion (e.g., Alexander, 1976; Gatesy

K.G. Hatala et al. / Journal of Human Evolution 112 (2017) 93e10494

et al., 1999; Wilson and Carrano, 1999; Ezquerra et al., 2007;Castanera et al., 2013).

Compared with their frequency in other paleontological records,the known sample of track sites from the early parts of the humanfossil record is sparse. Tracks attributed tomodernHomo sapiens areknown at sites all around the world from the late Pleistocene (e.g.,Mountain, 1966; Roberts and Berger, 1997; Webb et al., 2006; Kimet al., 2009; Liutkus-Pierce et al., 2016) and from the Holocene(e.g., Rector, 1979; Roberts et al., 1996; Meldrum, 2004;Mastrolorenzo et al., 2006; Morse et al., 2013). Sites earlier thanthese are much rarer. One famous site preserving Pliocene(~3.66 Ma; Deino, 2011) hominin trackways was discovered atLaetoli, Tanzania, in 1978 (Leakey andHay,1979). At the time of theirdiscovery, the oldest known hominin skeletal fossils were youngerthan 3.66 Ma, so these tracks provided indisputable evidence of theearliest known appearance of fossil hominins and also showed thatthese earliest known hominins were bipeds. These tracks are typi-cally assumed to have been produced by Australopithecus afarensis,which is still the only hominin known to have inhabited the Laetoliarea at this time (but see Tuttle et al., 1991). The Laetoli footprintsoffered direct evidence to support interpretations from fossil skel-etal morphology that early hominins with small brains and largeteeth walked bipedally (Leakey and Hay, 1979; Day and Wickens,1980; White, 1980; Leakey, 1981), definitively contradicting muchearlier hypotheses (e.g., Darwin, 1871). Around the same time (in1978), an assemblage of early Pleistocene (~1.4 Ma) hominin trackswas uncovered at Koobi Fora, Kenya (Behrensmeyer and Laporte,1981). Because that site was less extensive, involving only seventracks from one individual with variable preservation, and becauseof their younger age, these tracks from Koobi Fora have historicallyreceived less attention than those from Laetoli.

For two decades, these two were the only known hominin tracksites that predated the emergence of anatomicallymodern humans.As a result, despite an initial flurry of analyses regarding the Laetolitrackways, fossil hominin tracks have received considerably lessattention from paleoanthropologists than hominin skeletal fossils.Within the past 10e15 years, however, several sites that preservepre-H. sapiens (>315 ka; Hublin et al., 2017) tracks have beendiscovered in both Europe and Africa, and these have providedimportant new data for the human fossil record. In 2001, re-searchers conducted a detailed investigation of tracks long knownto exist on the slopes of Roccamonfina, a volcano in Italy. This studyconcluded that these tracks were between 325 and 385,000 yearsold and therefore must have been produced by a pre-H. sapienstaxon (Mietto et al., 2003; Avanzini et al., 2008). From2006 to 2008,multiple stratigraphic layers dating to ~1.5 Ma at Ileret, Kenya, werefound with preserved hominin trackways likely attributable toeitherHomo erectus sensu lato or Paranthropus boisei (Bennett et al.,2009). In 2013, a site at Happisburgh, UK, was uncovered by tidalerosion and found to contain hominin tracks dating to between0.78 and 1.0 Ma, preliminarily attributed to Homo antecessor(Ashton et al., 2014). Most recently, in 2015, two additional 3.66 Mahominin trackways were uncovered at Laetoli (Masao et al., 2016).This growing sample of hominin tracks from a variety of times andgeographical locations suggests that these new data, and additionaldiscoveries that may follow, can provide valuable contributionsthat will help us better understand our evolutionary history.

These discoveries also have sparked the development of newmethodological approaches for digitally recording fossil hominintrack assemblages (Bennett et al., 2009, 2013; Hatala et al., 2016a)and new experimental approaches for interpreting aspects of loco-motor biomechanics from hominin track morphologies (D’Aoûtet al., 2010; Raichlen et al., 2010; Crompton et al., 2011; Bates et al.,2013; Hatala et al., 2013, 2016a, b, c). These new techniques bypasscertain limitations of other paleontological and archaeological data

and can help to inform long-standing questions relating to theevolution of human anatomy, locomotion, and behavior.

Here, we present new evidence from multiple fossil hominintrack sites discovered in recent years within ~1.5 Ma fossil depositsnear Ileret, Kenya. In the years since the track surfaces at siteFwJj14E near Ileret were first described, expansions of these earlierexcavations combined with new surveys have led to discoveries ofnew tracks and trackways and multiple additional ~1.5 Ma tracksurfaces in nearby areas (Dingwall et al., 2013; Richmond et al.,2013; Hatala et al., 2016a; Roach et al., 2016, 2017). The totalassemblage of 1.5Ma hominin tracks known from the Ileret area hasgrown in size from the initially published sample of 20 tracks at asingle site (FwJj14E; Bennett et al., 2009) to a currently knownsample of 97 hominin tracks across five distinct sites (Hatala et al.,2016a). The size and nature of this assemblage of ~1.5 Ma hominintracks now permits the exploration of an entirely new series ofquestions that could not be addressed in the initial publication.Recently published analyses (Hatala et al., 2016a) have focused onthe implications of these tracks and trackways for locomotion andsocial behavior. We have also assessed the implications of thehominin tracks for patterns of land use and the overall paleoenvir-onmental context of the track surfaces (Roach et al., 2016, 2017). Inthe present study, we aim to 1) provide detail on the discoveries andexcavations of each of the five ~1.5 Ma hominin track sites at Ileret,including their geological and sedimentological contexts, to aidother researchers who may encounter or wish to search for similarsites, 2) analyze external dimensions of Ileret hominin trackswithina broad comparative framework, to further evaluate their taxo-nomic affiliation, 3) provide estimates of traveling speed for allexcavated Ileret trackways, and incorporate those results into pastinferences regarding social behavior, 4) explore, in detail, behavioralhypotheses consistent with the evidence found within the Ilerettrack sites, and 5) provide all raw data relevant to these analyses.

2. Discoveries and excavations of Ileret hominin tracksurfaces

2.1. Initial discoveries of hominin track surfaces at FwJj14E

The possibility that track surfaces might be preserved in Pleis-tocene sedimentary deposits near Ileret, Kenya was recognizedduring paleontological surveys (by A.K.B.) in the late 1970s, but itwas several decades before any tracks were discovered in this area.In 2005, paleontological excavations were started at site FwJj14Eafter the discovery of fossil skeletal material from a presumedP. boisei upper limb (Richmond et al., 2009). In order to understandthe geological/sedimentary context of that skeletal fossil discovery,a stepped trench was dug at the site of the subsequent excavation.While examining the stratigraphy of the site, Dr. Gail M. Ashleyrecognized that one sedimentary layer (later named the LowerFootprint Surface [LFS]) likely preserved animal tracks andmultipleother layers were also identified as possible track surfaces. In 2006,a section of the LFS was uncovered and found to preserve a largeassemblage of bovid tracks. In 2007, the excavation of the LFS wasexpanded, and new excavations were focused on a second potentialtrack surface (later named the Upper Footprint Surface [UFS]) thathad been identified at a higher (younger) position in the strati-graphic sequence. This initial excavation of the UFS in 2007revealed the first hominin tracks discovered at FwJj14E. Through2007 and 2008, the LFS and UFS were further excavated andanalyzed, along with several less extensive track surfaces betweenthese two layers. These initial discoveries and preliminary analysesof the track surfaces at FwJj14E were published in 2009 andrevealed a total of 20 hominin tracks across both the LFS and UFS(Bennett et al., 2009).

K.G. Hatala et al. / Journal of Human Evolution 112 (2017) 93e104 95

2.2. Continued excavations of hominin track surfaces at FwJj14E

From 2010 to 2014, excavations were continued at the site ofFwJj14E. Both the LFS and UFS were further exposed and severallayers between these two surfaces in the stratigraphic sequencewere identified as potential track surfaces (Fig. 1). Excavatorscleared small test squares (~1 m2) on several of these intermediatelayers to determine whether they might also preserve hominintracks. Through these continued excavations, a total of 53 addi-tional hominin tracks (48 on the UFS, three on the LFS, two on anintermediate layer named Layer A2) were uncovered at the site(Table 1).

These additional discoveries provide a wealth of new data andclarify earlier interpretations of the track surfaces (Fig. 2). Forinstance, it became clear that a trackway originally thought torepresent a single individual moving across the UFS (FUT1 sensuBennett et al., 2009) actually consisted of the trackways of twoindividuals, where one made prints that overlapped those ofanother individual (Dingwall et al., 2013; Richmond et al., 2013).The similar step lengths, and states of preservation of these track-ways, suggest that these individuals walked at similar speeds andmay have traveled across the surface at or around the same time. Italso became clear upon further excavation and detailed examina-tion that certain impressions formerly classified as potentialhominin tracks in the initial description of the site (FUI2, FUI5, FUI7within Bennett et al., 2009) could not be confidently attributed tohominins, although these impressions did not influence any pre-vious interpretations made by those authors. More detailed exca-vations also led previously unidentified impressions to berecognized as hominin tracks. In total, these ongoing excavationshave now exposed 54 m2 of the UFS, about 65 m2 of the LFS, andapproximately 2 m2 of the intermediate Layer A2. These surfaces atthe site of FwJj14E together preserve a total of 72 hominin tracks,including nine distinct sequences of tracks that have been identi-fied as continuous trackways (Table 1).

Figure 1. Photograph showing a cross-section of the sedimentary layers overlying the Fwthickness and not always continuous, which tend to be overlain by fine or silty sands. This dethe excavations at site FwJj14E, multiple bedded silt layers within the stratigraphic sequenLayer A2) preserved hominin tracks. Scale bar ¼ 10 cm.

2.3. Discoveries of sites ET-2013-1A-FE1 and ET-2013-1A-FE3

In 2013, expanded surveys for track-bearing sedimentary sur-faceswere conducted in East Turkana Collecting Area 1A, where siteFwJj14E is located. Paleontological Collecting Areas are delineatedby natural features such as rivers and were defined during theinitial rounds of research at Koobi Fora (Feibel, 2011). Surveys aimedto identify additional track surfaceswithin thewell-defined ~1.5MaIleret Tuff Complex (ITC), which is part of the Okote Member of theKoobi Fora Formation. The ITC is capped by the Northern Ileret Tuff,which is dated to 1.51e1.52 Ma, and it is bounded below by theLower Ileret Tuff, which is dated to 1.53 Ma (Brown et al., 2006;McDougall and Brown, 2006). At an intermediate position be-tween these two tuffs is the Ileret Tuff, which has been dated to1.52Ma (Bennett et al., 2009).Within the ITC is a ~8.5m sequence ofmassive laminated silts with intervening layers of fine-grained,stratified and cross-stratified sands that were likely deposited byintermittent low energy water transport processes within a deltamargin environment, near a lakeshore (Roach et al., 2016).

On June 28, 2013, three additional surfaces were identified whilesurveying the ITC exposures in Collecting Area 1A. The sedimentsoverlying these track surfaces had eroded near the edges, such thatportions of the track surface could be exposed simply by sweepingaway loose surface sand, and tracks were immediately visible. Twoof the surfaces discovered in this manner preserved hominin tracks.Upon discovery, and prior to excavation, these surfaces wereassigned field names of ET-2013-1A-FE1 and ET-2013-1A-FE3 (ET forEast Turkana, 2013 to designate the year of discovery, 1A to desig-nate the East Turkana Collecting Area, and FE for footprint excava-tion; these names are abbreviated hereafter as FE1 and FE3). FE1wasdiscovered by K.R.O. and FE3 by B.G.R. These surfaces were exca-vated in 2013, leading to the exposure of 8 m2 of the track surface atsite FE1 and 18 m2 of the track surface at site FE3. These excavationsrevealed many additional hominin tracks in situ, including fivetracks at site FE1 and 21 tracks at site FE3 (Figs. 3 and 4).

Jj14E LFS. Evident in this photograph are multiple layers of bedded silts, variable inpositional couplet is repeated multiple times between the FwJj14E LFS and UFS. Duringce were identified as having the potential to preserve tracks. One of these (designated

Table 1Catalog of hominin tracks discovered at site FwJj14E from 2007 to 2014.a

Track Trackway Tracksurface

Reported by Bennettet al. (2009)?

FLT1-1 FLT1 LFS XFLT1-2 FLT1 LFS XFLT1-3 FLT1 LFSFLT1-4 FLT1 LFSFLT1-6 FLT1 LFSFUI3 UFS XFU-A FU-O UFSFU-AA UFSFU-AB UFSFU-AC UFSFU-AD FU-AD UFSFU-AE FU-AD UFSFU-B FU-O UFSFU-C UFSFU-D UFSFU-E/FUI6 FU-E UFS XFU-F FU-E UFSFU-G FU-E UFSFU-H FU-E UFSFU-I FU-E UFSFU-J UFSFU-K FU-E UFSFU-L FU-O UFSFU-M FU-O UFSFU-N UFSFU-O FU-O UFSFU-P UFSFU-S UFSFU-T UFSFU-W UFSFU-X FU-X UFSFU-Y FU-X UFSFU-Z UFSFUT1-1 FUT1B UFS XFUT1-2 FUT1B UFS XFUT1-3 FUT1A UFS XFUT1-4 FUT1A UFS XFUT1-4i FUT1B UFSFUT1-4ii FUT1B UFSFUT1-5 FUT1A UFS XFUT1-5i FUT1B UFSFUT1-6 FUT1A UFS XFUT1-7A FUT1A UFS XFUT1-7B FUT1B UFS XFUT1-8 FUT1A UFSFUT1-8i FUT1B UFSFUT1-9 FUT1A UFSFUT1-10 FUT1A UFSFUT1-11 FUT1B UFSFUT1-12 FUT1A UFSFUT1-13 FUT1A UFSFUT1-14 FUT1A UFSFUT1-14i FUT1B UFSFUT1-15 FUT1A UFSFUT1-16 FUT1A UFSFUT1-17 FUT1B UFSFUT1-18 FUT1A UFSFUT1-18i FUT1B UFSFUT1-19 FUT1A UFSFUT1-20 FUT1B UFSFUT2–3 FUT2 UFSFUT2–2 FUT2 UFSFUT2–1 FUT2 UFSFUT2-0 FUT2 UFSFUT2-1 FUT2 UFS XFUT2-2 FUT2 UFS XFUT2-3 FUT2 UFS XFUT2-4 FUT2 UFS X

Table 1 (continued )

Track Trackway Tracksurface

Reported by Bennettet al. (2009)?

FUT3-1 FUT3 UFS XFUT3-2 FUT3 UFS XA2-H2 Layer A2A2-H3 Layer A2

a If a given track could be linked to a continuous trackway produced by the sameindividual, the field name of that associated trackway is listed. Indications areprovided if a given track was reported in the initial description of the site (Bennettet al., 2009).

K.G. Hatala et al. / Journal of Human Evolution 112 (2017) 93e10496

2.4. Discoveries of additional hominin track surfaces throughrandom spatial sampling

In 2014, surveys were conducted at randomly selected locationswithin the ITC exposures in Collecting Areas 1A and 3 as part of aprotocol designed to systematically (and in an unbiased fashion)sample track surfaces in order to assess the broader paleoecologicalcontext of the hominin track sites (Roach et al., 2016, 2017). Briefly, amap of the ITC exposures in these areas was gridded into sequen-tially numbered 20 m � 20 m squares using ArcGIS (v. 10.2). Arandom number generator (Microsoft Excel v. 14.4.8) was used tochoose 20 m � 20 m grid squares to survey for potential track sur-faces. If a surface with an identifiable track occurred at one of therandomly selected locations, then a 1 m � 1 m test square on thatsurfacewas excavated (Roach et al., 2016). Through this protocol, twoadditional hominin track surfaces were discovered. The field namesof these sites were ET-2014-3-FE8 and ET-2014-1A-FE16 (abbrevi-ated here as FE8 and FE16). K.R.O. and N.T.R. discovered site FE8 andK.G.H. discovered site FE16. The excavations at these sites are lessextensive than the others listed above because they were part of thesystematic footprint survey. FE8 consists of only a 1 m � 1 m square(Fig. 5). The FE16 excavation was slightly expanded when it wasrealized that a potential hominin track lay on the edge of the randomsquare and further excavation was needed to confirm its identity.That site consists of a 2 m � 1 m rectangle of exposed track surface(Fig. 6). At site FE8, one hominin track has been identified, and threehominin tracks have been identified at site FE16 (Figs. 5 and 6).

2.5. Preservation of data from track surfaces

Following direct measurements of tracks and trackways (seeMethods), the global three-dimensional positions of all tracks werelogged using a total station (Leica Builder 505; Trimble Nomad 900LE data collector with EDMce forWindowsMobile software). Entiretrack surfaces were then recorded using photogrammetry, as amethod of digitally quantifying but also preserving the originalpaleontological data (Falkingham, 2012). This method was deemedthe most useful approach for permanently recording the footprintsurfaces at Ileret, since they are located in a very remote location,preserved in unconsolidated sediment, and are at relatively highrisk of damage through natural erosion processes (Bennett et al.,2013). Further, these digital records of the footprint surfaces willbe curated as original records of the site that accompany totalstation data, field notes, and other paper records.

Even though track surfaces were digitally documented imme-diately following their excavation, considerable efforts were madeto rebury track surfaces in a manner that would aid their long-termpreservation. The sterile sand that had previously overlain the tracksurfaces and helped to preserve them for the past 1.5 million yearswas sieved and used to rebury each site. Nylon tarpaulins wereplaced on top of the sand layers and were subsequently secured byrocks. The underlying sand was graded in a manner that woulddirect water off the edge of the excavation site and prevent it from

Figure 2. Schematic map of site FwJj14E UFS. Figure is adapted from a previously published version (Hatala et al., 2016a), with copyright retained by K.G.H. The left and right imagesshow the extent of the excavations following the 2009 and 2014 field seasons, respectively. The solid red lines denote the excavation borders (i.e., they do not symbolize the limits ofthe surface, as the surface may continue beyond these). The dashed red line indicates the edge of the track surface and areas west of this line have been lost due to erosion of theoutcrop. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

K.G. Hatala et al. / Journal of Human Evolution 112 (2017) 93e104 97

pooling on top of the tarpaulin. The tarpaulins were subsequentlyburied with additional sterile, sieved sand excavated from thestratified layers that had overlain each track surface prior to exca-vation. These overlying sediments were also graded in a mannerthat would divert surface water flow away from the site of theexcavated track surface. Continuous work in the Koobi Fora areaeach year will allow for monitoring of these sites, and interventionwill be possible if any appear to be at risk of damage.

3. Geological/sedimentological contexts of Ileret hominintrack assemblages

Studies of the geological and sedimentological contexts of tracksurfaces have differed methodologically between the various tracksites. Because site FwJj14E has been the focus of continued study formore than a decade, the geology and sedimentology of this site is

Figure 3. Overhead view of a three-dimensional photogrammetric model of site ET-201continuous trackway are circled in solid white. Scale bar ¼ 1 m (displayed in the bottom leftthe top left corner.

understood in the greatest detail. Earlier interpretations of thestratigraphy at site FwJj14E (Bennett et al., 2009) have been fol-lowed bymore detailed reconstructions of the sedimentary contextof the FwJj14E track surfaces (Behrensmeyer, 2011; Roach et al.,2016). At the other hominin track sites at Ileret (FE1, FE3, FE8,and FE16), our understanding of site depositional processes isderived from detailed studies of the sedimentary layers andbedding structures exposed by the track surface excavations. Thesesites have also been tied in to the overall geology of the area andlinked to the stratigraphic sequence exposed at site FwJj14E bymeasuring relative positions within the sequence of three volcanictuffs (the Lower Ileret, Ileret, and Northern Ileret Tuffs) thatcomprise the ITC. It should be noted, however, that lateral variationin the track-bearing lithofacies makes it difficult to confirm theexact correlations of depositional surfaces at different sites. Hom-inin track surfaces were found onmultiple bedded silt layers within

3-1A-FE1. Hominin tracks are circled in dashed white. Hominin tracks that form acorner). The approximate direction of magnetic north is indicated by the black arrow in

Figure 4. Overhead view of a three-dimensional photogrammetric model of site ET-2013-1A-FE3. Hominin tracks are circled in dashed white. Scale bar ¼ 1 m (displayed on the leftside of the image). The approximate direction of magnetic north is indicated by the black arrow in the top left corner.

Figure 5. Overhead view of a three-dimensional photogrammetric model of site ET-2014-3-FE8. The single hominin track on this surface is circled in dashed white. Scalebar ¼ 1 m (shown on the left side of the image). The approximate direction of magneticnorth is indicated by the black arrow in the top left corner.

Figure 6. Overhead view of a three-dimensional photogrammetric model of site ET-2014-1A-FE16. The three definitive hominin tracks discovered on this surface arecircled in dashed white. Because of the small size of this excavation, it is unclear at thispoint whether these tracks form a continuous trackway. As a result, they have not beenassigned to a trackway. Scale bar ¼ 1 m (shown on the left side of the image). Theapproximate direction of magnetic north is indicated by the black arrow in the top leftcorner.

K.G. Hatala et al. / Journal of Human Evolution 112 (2017) 93e10498

the ITC, with one of these surfaces (the FwJj14E UFS) lying betweenthe Ileret and Northern Ileret Tuffs and the other six (FwJj14E LFS,FwJj14E Layer A2, FE1, FE3, FE8, FE16) lying between the LowerIleret and Ileret Tuffs (Fig. 7).

The hominin track surfaces were buried by fine silty sand aftertracks were produced on the bedded silt layers, and this likelyoccurred very rapidly. The lack of mud cracking on any of thesesurfaces is consistentwith ahigh, stablewater table, and theabsenceof root traces or other evidence of pedogenesis in track-bearingsediments indicates that they were not sub-aerially exposed(Roach et al., 2016). Paleosols also occur within the ITC and provideevidence for the temporary development of stable land surfaces, butthe intervening periods of low energy deposition of interbedded siltand sand on the margin of a delta or lake allowed tracks to be pro-duced and track surfaces to be preserved repeatedly in this areaduring the approximately 20 ka spanned by the ITC (Fig. 7).

4. Methods

4.1. Measurement and documentation of fossil hominin tracks andtrackways

Following excavation and exposure of track surfaces, multipletechniques were used to measure and record the track andtrackway data. First, the linear dimensions of individual tracks(lengths and widths) were measured directly. The exact linearmeasurements taken from each track were dependent upon thespecific nature of its preservation. Some tracks represented onlyparts of the foot, and in some cases, the anatomical definition of thetrack was somewhat distorted. As such, multiple measurements oftrack length and breadth were attempted. A tapemeasurewas usedto measure the track's length along one or more (where possible) ofthree different trajectoriesdfrom the most proximal part of theoutline of the heel impression to the tips of the impressions for thehallux and/or second and/or third digits. In some tracks, the im-pressions for certain digits were unclear and the measurementswere estimated (and recorded as such) or excluded. Track breadthswere measured using digital calipers at two different locations. Thefirst measurement was taken between the impressions created bythe first and fifth metatarsal heads, and the second was takenacross the widest part of the heel impression.

Based on visual observations of track locations and track mor-phologies, and consideration of individual track dimensions,certain tracks could immediately be linked to others within track-ways consisting of multiple steps by the same individual. In these

Figure 7. Representative stratigraphic section showing the positions of Ileret hominintrack surfaces within the ITC. Figure is derived from a previously published version(Roach et al., 2016), with copyright retained by N.T.R. Scale at bottom indicatesapproximate grain sizes: C ¼ clay, Z ¼ silt, S ¼ sand, G ¼ gravel. One surface, theFwJj14E UFS, lies between the Ileret and Northern Ileret Tuffs, while the other sixsurfaces lie between the Lower Ileret and Ileret Tuffs. The depositional contexts of eachof these surfaces is similar, as they lie within stratified and interbedded silt and fine-grain sand layers deposited by low energy processes, probably within a delta or lakemargin. These well-bedded intervals are separated by paleosols representing inter-vening periods of subaerial delta plain (fluvial) deposition.

K.G. Hatala et al. / Journal of Human Evolution 112 (2017) 93e104 99

cases, where trackways were evident, step and/or stride lengthswere directly measured with a tape measure. A field name wasassigned to each trackway after its identification.

Following direct measurements, hundreds of digital photo-graphs were taken in order to render high-resolution photogram-metric 3D models of entire track surfaces and all individualhominin tracks. These 3D records are important for long-term datapreservation (see above), but high resolution 3Dmodels of hominintracks have also enabled quantitative comparative analyses withtracks of other fossil and modern taxa (Hatala et al., 2016a).

4.2. Analyses of fossil hominin tracks and trackways

4.2.1. Analyses of external track dimensions Linear measurementsof hominin tracks from the Ileret track surfaces were comparedwith a compilation of similar measurements from a variety ofextant and extinct samples. Comparative data from modern taxaincluded measurements of tracks from footprint formation exper-iments conducted by K.G.H. with habitually barefoot Daasanachpeople from near Ileret, Kenya (Hatala et al., 2016b) and withmodern chimpanzees in the Primate Locomotion Lab at StonyBrook University (Hatala et al., 2016c). Fossil data consisted of acompilation of published linear measurements from other fossilhominin track sites, including 325e385 ka tracks fromRoccamonfina, Italy (Avanzini et al., 2008), 0.78e1.0 Ma tracksfrom Happisburgh, UK (Ashton et al., 2014), and ~3.66 Ma tracksfrom Laetoli, Tanzania (Bennett et al., 2016; Hatala et al., 2016c;Masao et al., 2016).4.2.2. Estimates of traveling speed After taking linear measure-ments of stride lengths (or in some cases step lengths), we

estimated traveling speeds from each of the Ileret hominin track-ways. To do so, we used a linear regression relationship betweenstride length and speed that was derived in an earlier humanexperimental study (speed ¼ �1.39 þ (0.48*(stride length/averagefootprint length)); Dingwall et al., 2013). Traveling speed estimatesfor some of these trackways have been published previously(Dingwall et al., 2013), but the estimates that we present hereincorporate new data (including newly excavated tracks and newmeasurements of stride lengths) that were obtained during ourexpansions of the site excavation.

5. Results

5.1. Analyses of external track dimensions

Themeasurementsof the lengthsandbreadthsof each track in theIleret assemblage are provided in Supplementary Online Material(SOM) Table S1. Comparisons of heel to hallux length and breadthacross the forefoot among the Ileret track surfaces and other fossilhominin track sites are given in Table 2. The tracks from each of the~1.5 Ma Ileret track surfaces are generally comparable in size to thetracks produced in footprint formation experiments by modernDaasanach people (Hatala et al., 2016a, b) and to the 325e385 katrackspreservedatRoccamonfina.Theyare longerandnarrower thana collection of tracks produced experimentally by modern chim-panzees. On average, the Ileret tracks are larger than those preservedat the 0.78e1.0Ma site of Happisburgh. They are generally longer butsimilarly wide to the ~3.66 Ma tracks from Laetoli, Tanzania.

5.2. Estimates of traveling speeds

Most of the identified trackways, and hence most of these trav-eling speed estimates (eight of 10), come from our most expansiveexcavations on the LFS and UFS at site FwJj14E. All of the 10 track-ways that we have identified almost certainly represent walkingspeeds, with estimates ranging from 0.45 to 1.58 m/s (Table 3). Oneisolated trackway on the LFS at FwJj14E had been thought torepresent a speed within the range at which modern Daasanachpeople tend to transition from a walk to a run (2.0e2.3 m/s;Dingwall et al., 2013), but our revised estimate here falls squarely inthe observed range of modern human walking speeds.

6. Discussion

6.1. Implications of analyses of external track dimensions

The Ileret tracks are human-like in their external sizes. Anydifferences between these and more recent hominin tracks arealmost certainly attributable to differences in the demographics ofthe hominins who produced them. For example, at Happisburghthere are many smaller tracks that are hypothesized to have beenproduced by children (Ashton et al., 2014). The larger Happisburghtracks, presumably produced by adults, are generally comparable insize to the tracks at Ileret.

Foot skeletal fossils are known for the two taxa proposed asmostlikely responsible for the Roccamonfina and Happisburgh tracks e

Homo heidelbergensis and Homo antecessor, respectively. Analyses ofthese H. heidelbergensis and H. antecessor foot bones have suggestedoverall foot sizes similar to those of modern humans (Lu et al., 2011;Pablos et al., 2015). Here, our results suggest that modern human-like foot sizes may have extended back even further in the homi-nin clade, to at least the time of the Ileret tracks at ~1.5 Ma.

The ~3.66 Ma tracks from Laetoli, Tanzania are approximately aswide as the Ileret tracks, but generally are not as long. The averagelength of tracks from one Laetoli trackway exceed the pooled

Table 2Linear dimensions of the tracks from various Ileret hominin track surfaces and multiple comparative samples.a

Sample Footprint length (cm) Footprint breadth (cm)

Mean Range SD Mean Range SD

Modern human (Daasanach) 25.4 (n ¼ 41) 20.0e29.5 2.1 9.7 (n ¼ 41) 7.4e11.8 0.9Modern chimpanzeesb 20.3 (n ¼ 2) 19.9e20.7 0.6 11.3 (n ¼ 2) 11.0e11.6 0.5Roccamonfinac 24 (n ¼ ?) e e 12 (n ¼ ?) e e

Happisburghd 19.1 (n ¼ 12) 14.0e26.0 4.2 7.5 (n ¼ 12) 5.0e11.0 1.8Laetolie 22.8 (n ¼ 4) 19.1e26.1 2.9 9.6 (n ¼ 2) 8.7e10.4 1.2Ileret (all sites pooled) 25.3 (n ¼ 28) 20.5e30.5 2.2 9.6 (n ¼ 36) 7.5e12.7 1.2Ileret FwJj14E UFS 25.2 (n ¼ 18) 20.5e30.5 2.0 9.8 (n ¼ 19) 8.25e12.7 1.4Ileret FE1 23.6 (n ¼ 3) 21.0e26.8 2.9 8.0 (n ¼ 3) 7.5e9.0 0.9Ileret FE3 26.3 (n ¼ 5) 23.5e29.5 2.9 9.9 (n ¼ 12) 7.5e12.0 1.2Ileret FE8 27.0 (n ¼ 1) e e 9.5 (n ¼ 1) e e

Ileret FwJj14E LFS 24.9 (n ¼ 1) e e 8.0 (n ¼ 1) e e

a The comparative samples include experimentally produced tracks made by modern humans and chimpanzees, and fossil tracks from the sites of Roccamonfina(325e385 ka), Happisburgh (0.78e1.0 Ma), and Laetoli (~3.66 Ma). For each sample, the average footprint dimensions are calculated by first computing means within eachtrackway (i.e., for each individual), and then using those trackway/individual means to calculate the average across the entire sample. SD ¼ standard deviation.

b These measurements represent unpublished data from K.G.H. The experiments in which these data were produced are described in Hatala et al. (2016a).c Data are from Avanzini et al. (2008). Sample size is listed as a question mark because, although 56 tracks are described to exist at the site, it is unclear exactly how many

were measured to arrive at the average measurements included in the publication.d Data are from Ashton et al. (2014).e Data included from four Laetoli trackways e G1, G3, S1, and S2. Data for Laetoli G1 were collected by K.G.H. Measurements of Laetoli G3 are from Bennett et al. (2016).

Laetoli S1 and S2 data are from Masao et al. (2016). Average lengths from each of the four trackways were used to derive the ‘population’ average footprint length, butconfident width measurements were only available for the G1 and S1 trackways.

Table 3Estimates of traveling speed for Ileret hominin trackways.a

Trackway Surface Step length (cm) Stride length (cm) Estimated speed (m/s)

FLT1 FwJj14E LFS 110.0 0.73FUT1A FwJj14E UFS 86.5 0.65FUT1B FwJj14E UFS 73.3 0.45FUT2 FwJj14E UFS 124 1.05FUT3 FwJj14E UFS 130 1.23FU-E FwJj14E UFS 156.5 1.52FU-O FwJj14E UFS 131.5 1.32FU-X FwJj14E UFS 114.0 1.29FU-AD FwJj14E UFS 83.3 1.58FE1-HT1 FE1 42.9 0.58

a Step and stride lengths are provided. Traveling speed estimates were produced using the equation provided by Dingwall et al. (2013). In cases where a stride lengthmeasurement was not possible, stride length was estimated as two times step length.

2 We acknowledge here that these previously published body mass predictionsrely upon a reference sample of modern humans who have a very linear buildcompared with other modern populations (Ruff, 1994). If H. erectus were charac-terized by a relatively wide pelvis and a less linear build (Ruff, 2010), our experi-mental design may have produced underestimates of body mass for H. erectusindividuals (Ruff and Burgess, 2015).

K.G. Hatala et al. / Journal of Human Evolution 112 (2017) 93e104100

average track length from Ileret (individual S1, average tracklength ¼ 26.1 cm; Masao et al., 2016) but the average lengths of allother Laetoli trackways fall below this average (average lengths ofG1, G3, S2 range from 19.1 to 23.1 cm; Bennett et al., 2016; Hatalaet al., 2016c; Masao et al., 2016). It has been proposed that theLaetoli track assemblage represents considerable body size varia-tion, and possibly a high degree of sexual dimorphism in overall sizewith trackway S1 representing a large male (Masao et al., 2016).Assuming that track size is an accurate predictor of overall body sizefor fossil hominins (e.g., Dingwall et al., 2013; Hatala et al., 2016a),our results here show that the Laetoli track makers were still, onaverage, smaller in overall size than those from Ileret. This resultagrees well with a recent analysis by Grabowski et al. (2015), whichpredicted fossil hominin body masses from lower limb skeletalfossils. That analysis included a range of bodymass predictions fromfossils attributed to both A. afarensis, the taxon most commonlyassumed to have created the Laetoli tracks (e.g., Masao et al., 2016but see; Tuttle et al., 1991), and H. erectus, the presumed maker ofthe Ileret tracks (Hatala et al., 2016a). Grabowski et al. (2015) foundthatwhilemale A. afarensis individualsmay have had body sizes thatfell within the range observed in H. erectus, the species was still, onaverage, smaller than H. erectus.

These comparisons with other fossil skeletal and track data rely,to some extent, upon accurate taxonomic attribution of the Ilerethominin tracks. Using the relationships between track size and

body size determined through experiments with modern Daasa-nach people, body mass predictions were generated for many Ilerethominin tracks and trackways (Hatala et al., 2016a). These bodymass estimates from the Ileret hominin tracks were largely similarto the observed body masses of modern Daasanach people, andcloser to skeletally based estimates from H. erectus fossils than theywere to estimates derived from confidently attributable P. boisei orHomo habilis skeletal fossils (Grabowski et al., 2015). This line ofevidence was used to support attribution of the tracks to H. erectus(Hatala et al., 2016a).2 The data in Table 2 demonstrate that large,human-like track size is consistent across all of the excavated Ilerettrack surfaces. While some Ileret track surfaces include relativelysmaller tracks, none fall outside of the range of sizes observed intracks made by adult modern humans. Because we have notobserved differences in external track dimensions (this study) orinternal track morphologies (Hatala et al., 2016a) between Ilerettrack sites, or between the large and small tracks within the as-semblages, it is most parsimonious for us to hypothesize that all of

K.G. Hatala et al. / Journal of Human Evolution 112 (2017) 93e104 101

the Ileret tracks were made by one hominin taxon and that taxonwasmost likelyH. erectus. However, at this point we cannot rule outcompletely that another hominin species made some or all of thetracks: a paucity of postcranial skeletal fossils means we know verylittle about body size variation in P. boisei or H. habilis, two taxa thatare known to have co-existed with H. erectus in the Ileret areaaround ~1.5 Ma, and even less about Homo rudolfensis.

6.2. Implications of estimates of traveling speeds

Roach et al. (2016) evaluated whether certain Ileret hominintrackways on the FwJj14E UFS may have represented a groupcoordinating their movement and traveling together. To do so, theyused total station data to quantitatively compare the compass ori-entations of hominin and non-hominin trackways. They found thatthe hominin tracks were predominantly and non-randomly ori-ented in a southeasterly direction, and the tracks of all other ani-mals had a significantly different northwesterly orientation (Roachet al., 2016). These different directions of travel for the homininsand all other animals suggest that the movement patterns capturedon the FwJj14E UFS were not constrained by features of the naturallandscape, as can be observed when animals follow well-traveledgame trails (e.g., Laporte and Behrensmeyer, 1980). Further, withthe exception of two trackways (FUT1A and FUT1B) that overlap ina manner consistent with one individual following the other, theUFS preserves many sub-parallel hominin tracks that do not over-lap with one another. Roach et al. (2016) hypothesized that thesenon-overlapping, sub-parallel tracks could represent multiple in-dividuals simultaneously moving through this area together.

Here, we are able to use traveling speeds estimated from theFwJj14E UFS trackways, combined with published data on hominintrackway orientations, in order to further evaluate the likelihoodthat particular trackways may represent groups of individualsmoving together across the FwJj14E UFS. Among the trackwaysdescribed in Table 3, for which traveling speeds could be estimated,six of the eight (FUT1A, FUT1B, FUT3, FU-E, FU-O, and FU-X) areoriented in a southeasterly directionwhile the other two (FUT2 andFU-AD) are oriented towards the northwest. These trackways sug-gest a range of walking speeds that would not preclude grouptravel, as individuals within a group may vary their speeds whilestill traveling together. It is notable that the FU-X individual createdthe smallest tracks within the entire Ileret sample (SOM Table S1),and in a previously published analysis, Hatala et al. (2016a) foundthat this trackway generated a body mass estimate that was anextreme outlier among the Ileret sample, being far smaller thanaverage (26.5 kg). In this study, we estimate that this markedlysmaller individual still traveled at a speed consistent with the otherindividuals who moved towards the southeast, perhaps making aneffort to keep up with the rest of the group. At this point, it is stilldifficult to hypothesize exactly who may have traveled togetheracross the FwJj14E UFS, but trackway orientations and travelingspeeds estimates appear to be consistent with what would be ex-pected from coordinated group movement.

6.3. Novel directions for testing hypotheses related to homininpaleobiology at ~1.5 Ma

Evolutionary hypotheses related to early hominin groupbehavior and social structure have proven notoriously difficult totest due to inherent limitations of the most common types of hu-man fossil data (Chapais, 2013). While certain aspects of skeletalmorphology such as canine size dimorphism may be associatedwith broad patterns of social behavior in primates (Leuteneggerand Shell, 1987), those links are often tenuous and not broadlyapplicable across diverse taxa (Plavcan, 2000). The Ileret track

assemblages, however, represent immediate snapshots of fossilhominin behaviors and they are recorded at a level of spatiotem-poral resolution that has the potential to directly inform hypothe-ses about group structures and interactions between fossil homininindividuals.

The sedimentological contexts of each of the Ileret hominin tracksurfaces are consistent with a short duration of surface exposure,during which tracks and trackways were formed and then rapidlyburied. Each hominin track surface was buried by water-transported, fine silty sand (Figs. 1 and 7) that infilled the trackswithout scouring or erosion. Similar low energy cycles of silt andsand deposition occur today along the shoreline of Lake Turkana,suggesting similar circumstances during the time of the ITC in adelta or lake margin environment. Taphonomic experiments on thedurations ofmodern human tracks and trackways along the shore ofmodern Lake Turkana suggest that, when left unburied, detailedmorphologies within the tracks (as are seen in the fossil hominintrack assemblages) normally have a lifespan of 1.3 days or less(Roach et al., 2016). Beyond this point, track morphologies becomeless well defined and less recognizable, typically due to weatheringor trampling by other animals (Roach et al., 2016). Together, theselines of evidence suggest that the Ileret hominin track surfaceswerecreated and buried within a narrow time span from a few days to afew hours. This very short timeframe of site formation and depo-sition suggests that any individuals or animals that made tracks onthe same surface likely lived within sufficiently close proximity toeach other that their ranges overlapped on a daily basis.

On the FwJj14E UFS, we find multiple sub-parallel hominintrackways, traveling in a similar direction that is statisticallydistinct from the directions of travel of all other animal tracks andtrackways (Roach et al., 2016). Coupled with the short windows oftime during which all of the trackways were formed and buried,these data suggest that at least some of the FwJj14E UFS individualsmoving in the same direction traveled together. Here, our estimatesof traveling speeds for the FwJj14E UFS trackways, which all implywalking speeds, provide additional support for the hypothesis ofcoordinated group movement. Even if these individuals happenedto move through the area at similar speeds but at slightly differenttimes (if they did not travel as a group), the very limited time forsite formation and burial evident from geological and sedimento-logical data means that the different individuals whose tracks arepreserved on the same surface lived in immediate proximity andlikely interacted with each other.

Hatala et al. (2016a) used body mass estimates derived fromexternal track dimensions in order to estimate the sexes of theindividuals who created the two largest assemblages of tracks thatwe have excavated at Ileret. They estimated that at least 50% of thehominin trackways on the FwJj14E UFS, and at least 75% of thetracks at site FE3 were created by male H. erectus individuals(Hatala et al., 2016a). Based on the results of the current study, wefurther emphasize that it is not necessarily the case that the Ilerettrackways were made by groups of exclusively males movingtogether. In fact, the group of sub-parallel southeast-orientedtrackways on the FwJj14E UFS, which also imply similar walkingspeeds and therefore offer the strongest case for coordinated groupmovement, includes a set of trackways that were all estimated byHatala et al. (2016a) as potentially representing females and/orsubadults. But while a small group of females and/or subadults mayhave moved together, they represent only six of the estimated 18different individuals who left tracks on the FwJj14E UFS (Table 2).What we find most striking and well-supported regarding socialbehavior is that within both the FwJj14E UFS and the FE3 trackassemblages, there are several trackways with sizes that exceed thepooled Ileret mean (Table 2) and likely represent multiple maleH. erectus individuals.

K.G. Hatala et al. / Journal of Human Evolution 112 (2017) 93e104102

The presence and presumed interactions of multiple H. erectusmales within the same local group has interesting implications thatare relevant to major evolutionary hypotheses. Direct tests of hy-pothesized social behaviors have been particularly difficult toachieve due to a lack of pertinent data from the human fossil re-cord. While there are certainly limitations to the data preserved inthese track assemblagesdfor example, the exact types of associa-tions or genetic relationships between individuals and the activitiesthat led to their interactions will never be knowndwe can examinethe evidence of group composition and look for consistencies with,or deviations from, existing hypotheses related to the evolution ofhuman social structure. We present below a series of behavioralhypotheses, each of which could be supported by evidence from theIleret track assemblages. This is not an exhaustive list of possibil-ities, but provides multiple scenarios that can be explored in futurework to refine methods for inferring behavior from assemblages oftracks and trackways.

First, it is possible that the Ileret track assemblages were pro-duced by group behavior patterns that have deep evolutionaryroots and may have characterized the Pan-Homo last commonancestor (LCA). Among modern chimpanzees, multi-male groupsare known to assemble and travel together while conducting pa-trols of their borders as a method of mitigating intergroupcompetition (Watts and Mitani, 2001; Mitani and Watts, 2005;Watts et al., 2006). A similar type of behavior may be evidencedby the large assemblages of presumed male tracks at the FwJj14EUFS and site FE3. Many of the hominin trackways on the FwJj14EUFS appear to represent individuals moving along the water's edge(Roach et al., 2016), which may or may not be consistent with agroup actively surveying a territorial perimeter. Multi-male groupsof chimpanzees are also known to engage in cooperative huntingexpeditions, although this behavior may not be generalizableacross the species as a whole because it has been observed at only asingle study site (Boesch, 2002). At this point, we do not havesufficient modern or fossil data to rule out the possibility that theIleret track assemblages could represent a behavior analogous tothat observed in modern chimpanzees. Future analyses of howspecific chimpanzee behaviors are recorded in tracks and track-ways are necessary to test these hypotheses.

The Ileret track assemblages do meet the predictions of themore general hypothesis that multi-male groups with some levelof cooperation were present in the LCA (Wrangham and Pilbeam,2001; Duda and Zrzavy, 2013). Yet, in order for human andchimpanzee maleemale cooperation to be homologous, thisstrategy must have been inherited from the LCA and persisted inthe lineages leading to both modern humans and chimpanzees.The high levels of size dimorphism in A. afarensis (Plavcan et al.,2005; Gordon et al., 2008; but see also Reno and Lovejoy, 2015),a species widely regarded as a stem hominin (Kimbel andDelezene, 2009), suggest high levels of maleemale competition.This hypothesized competition, plus little to no information on theexistence of cooperative behaviors such as group defense and/orhunting in A. afarensis, do not provide substantial support for thehomology hypothesis at this time. However, given the growingnumber of contemporaneous taxa recognized in the Pliocene(Haile-Selassie et al., 2016), the poor preservation of many ofthem, and the inherent difficulties in identifying ances-toredescendant relationships in the fossil record, it is prematureto rule out the hypothesis that the maleemale cooperationobserved in modern chimpanzees and humans may be homolo-gous. At the very least, the evidence from the Ileret track surfacesof multiple H. erectus males walking across the same landscape,and possibly even traveling together, is consistent with a level ofmaleemale cooperation similar to that observed in modernchimpanzees.

Alternatively, the Ileret track assemblages may preserve directevidence of derived behaviors that led to the emergence of a socialstructure that, among living primates, uniquely characterizesmodern humans. Maleemale cooperation is a key feature thatunderlies the maintenance of multilevel social structures, whichdescribe nearly all modern human societies and have been hy-pothesized, based on a variety of other archaeological and pale-ontological evidence, to have first emerged in H. erectus (Swedelland Plummer, 2012). This model suggests a tiered hierarchy ofdifferently sized social unitsdincluding, from smallest to largest,the polygynous one-male unit, the clan, the band, and thetroopdand coordination at all but the lowest levels of this hierar-chy depends upon, at the very least, mutual tolerance among males(Swedell and Plummer, 2012). Among modern human hunter-gatherers, it has been demonstrated that this nested, hierarchicalpattern of social structure may represent an ‘optimal’ solution forthe distribution of resources both among and within groups(Hamilton et al., 2007). As H. erectus emerged and evolved within adynamically changing environment that required wider dispersaland provided less reliable resources (Potts, 1998; DeMenocal, 2004;Ant�on et al., 2014; Potts and Faith, 2015), this social structure couldhave played an essential adaptive role in promoting foraging suc-cess. In modern human hunter-gatherers, multiple males fromwithin the same band cooperate to hunt for game and then sharethe calorically rich profits of that hunt among all members of theirband (Hill, 2002; Gurven, 2004; Marlowe, 2005; Hill et al., 2009).Cooperative foraging and the sharing of resources seems likely ifH. erectus regularly foraged for animal resources, as has been sug-gested by collections of cut-marked bone from 1.5 Ma deposits inCollecting Area 1A (Pobiner et al., 2008) and potentially by theIleret tracks themselves (Roach et al., 2017). Cooperation improvesthe probability of a successful kill in modern hunter-gatherers andmay have been evenmore important at the time of H. erectus, whenweapons were likely much less lethal than those used by hunter-gatherers today (Hill, 2002). Further, the sharing of food re-sources acquired through cooperative hunting across all membersof the band buffers the relatively high probability of failure that isassociated with any given hunt (Lee, 1968; Hill, 2002; Gurven,2004). If a multilevel social structure was present in H. erectus,mediated by maleemale cooperation and perhaps including pat-terns of cooperative foraging, then this structure could have pro-moted the emergence of the number of unique behavioral traitsand the “deep social structure” that are considered to definemodern humans (Hill et al., 2009; Chapais, 2011).

Ultimately, comprehensive tests of these types of hypothesesabout social behavior and social structure require more modernand fossil data than are currently available. Determining the pres-ence or absence of specific patterns of social behavior will rely uponcontinued work to refine our abilities and better understand ourlimitations for testing hypotheses about hominin social interactionsfrom track and trackway data. Experimental research is necessaryin order to determine how behaviors of humans, and of nonhumanprimates, are recorded in assemblages of their tracks. Thecontinued pursuit of such methods is valuable, as tracks andtrackways offer unique opportunities to directly observe, in deeptime, snapshots of groups of our fossil relatives and draw inferencesregarding the compositions and social behaviors of those groups.Track sites may allow for types of inferences regarding groupbehavior that have been notoriously difficult to gain through otherforms of archaeological and paleontological data (Chapais, 2013).Already, we have been able to draw some conservative conclusionsfrom the Ileret track assemblages. Regardless of the specifics of thesocial structure and social behaviors of H. erectus, our analyses dosuggest the presence and interactions of multiple H. erectus malesliving and traveling on the same landscapes. This evidence is

K.G. Hatala et al. / Journal of Human Evolution 112 (2017) 93e104 103

consistent with previous hypotheses that have suggested that thesocial structure of H. erectus supported the emergence of modernhuman-like patterns of social behavior.

7. Conclusions

These discoveries of multiple hominin track sites in OkoteMember deposits near Ileret, Kenya, provide a unique window thatcan further our knowledge of various aspects of hominin paleobi-ology at ~1.5 Ma. Initial analyses of these sites explored the impli-cations of these track sites for hominin anatomy, locomotion, landuse patterns, and behavior (Bennett et al., 2009; Hatala et al.,2016a; Roach et al., 2016, 2017). Here, we provide new compara-tive assessments of external track dimensions, and estimates oftraveling speed derived from hominin trackways, to refine earlierinterpretations of foot size, taxonomic attribution, and groupmovement. These new analyses, combined with those publishedpreviously, highlight the utility of trace fossil data in developingand testingmajor hypotheses regarding the biology and behavior offossil hominins. We hope that the details we provide on how wediscovered and excavated the Ileret hominin track sites, our de-scriptions of the geological/sedimentological contexts in which wefind these sites, and the initial analyses and hypotheses laid out inthis study will help motivate continued research in hominin ich-nology. Continued discoveries of new hominin trace fossil data, andthe development of new approaches for interpreting them, willhelp us use these data in concert with other parts of the paleon-tological and archaeological records to better inform un-derstandings of our evolutionary past.

Acknowledgements

We thank Ren�e Bobe, Andrew Du, Matt Ferry, Purity Kiura,Emma Mbua, Emmanuel Ndiema, Jonathan Reeves, Erin MarieWilliams-Hatala, students of the Koobi Fora Field School, the Na-tional Museums of Kenya, the town of Ileret, Kenya, and the localDaasanach volunteers for their contributions to this research. Thisstudy was conducted under a research permit granted by theKenyan National Council for Science and Technology and an exca-vation license granted by the Ministry of Higher Education, Scienceand Technology. This study was funded by the Leakey Foundation,the National Science Foundation (BCS-1232522, BCS-0924476, BCS-1128170, BCS-1515054, BCS-0935321, DGE-080163, SMA-1409612),the Wenner-Gren Foundation (Grant 8592), and The GeorgeWashington University's Research Enhancement Fund.

Supplementary Online Material

Supplementary online material related to this article can befound at http://dx.doi.org/10.1016/j.jhevol.2017.08.013.

References

Alexander, R.M., 1976. Estimates of speeds of dinosaurs. Nature 261, 129e130.Ant�on, S.C., Potts, R., Aiello, L.C., 2014. Evolution of early Homo: An integrated

biological perspective. Science 345, 45.Ashton, N., Lewis, S.G., De Groote, I., Duffy, S.M., Bates, M., Bates, R., Hoare, P.,

Lewis, M., Parfitt, S.A., Peglar, S., Williams, C., Stringer, C., 2014. Hominin foot-prints from early Pleistocene deposits at Happisburgh, UK. PLoS One 9, e88329.

Avanzini, M., Mietto, P., Panarello, A., De Angelis, M., Rolandi, G., 2008. The devil'strails: Middle Pleistocene human footprints preserved in a volcanoclastic de-posit of southern Italy. Ichnos 15, 179e189.

Bates, K.T., Savage, R., Pataky, T.C., Morse, S.A., Webster, E., Falkingham, P.L., Ren, L.,Qian, Z., Collins, D., Bennett, M.R., McClymont, J., Crompton, R.H., 2013. Doesfootprint depth correlate with foot motion and pressure? J. R. Soc. Interface 10,20130009.

Behrensmeyer, A.K., 2011. Conversations with Glynn's ghost: The evolution ofpaleolandscape research at East Turkana. In: Sept, J., Pilbeam, D. (Eds.), Casting

the Net Wide: Papers in Honor of Glynn Isaac and His Approach to HumanOrigins Research. Oxbow Books, Oakville, pp. 21e39.

Behrensmeyer, A.K., Laporte, L., 1981. Footprints of a Pleistocene hominid innorthern Kenya. Nature 289, 167e169.

Bennett, M.R., Harris, J.W., Richmond, B.G., Braun, D.R., Mbua, E., Kiura, P., Olago, D.,Kibunjia, M., Omuombo, C., Behrensmeyer, A.K., Huddart, D., Gonzalez, S., 2009.Early hominin foot morphology based on 1.5-million-year-old footprints fromIleret, Kenya. Science 323, 1197e1201.

Bennett, M.R., Falkingham, P., Morse, S.A., Bates, K., Crompton, R.H., 2013. Preservingthe impossible: conservation of soft-sediment hominin footprint sites andstrategies for three-dimensional digital data capture. PLoS One 8, e60755.

Bennett, M.R., Reynolds, S.C., Morse, S.A., Budka, M., 2016. Laetoli's lost tracks: 3Dgenerated mean shape and missing footprints. Sci. Rep. 6, 21916.

Bibi, F., Kraatz, B., Craig, N., Beech, M., Schuster, M., Hill, A., 2012. Early evidence forcomplex social structure in Proboscidea from a late Miocene trackway site inthe United Arab Emirates. Biol. Lett. 8, 670e673.

Boesch, C., 2002. Cooperative hunting roles among Tai chimpanzees. Hum. Nat. 13,27e46.

Brown, F.H., Haileab, B., McDougall, I., 2006. Sequence of tuffs between the KBS Tuffand the Chari Tuff in the Turkana Basin, Kenya and Ethiopia. J. Geol. Soc. Lond.163, 185e204.

Castanera, D., Vila, B., Razzolini, N.L., Falkingham, P.L., Canudo, J.I., Manning, P.L.,Galobart, �A., 2013. Manus track preservation bias as a key factor for assessingtrackmaker identity and quadrupedalism in basal ornithopods. PLoS One 8,e54177.

Chapais, B., 2011. The deep social structure of humankind. Science 331, 1276e1277.Chapais, B., 2013. Monogamy, strongly bonded groups, and the evolution of human

social structure. Evol. Anthropol. 22, 52e65.Cotton, W.D., Cotton, J.E., Hunt, A.P., 1998. Evidence for social behavior in ornith-

opod dinosaurs from the Dakota group of northeastern New Mexico, U.S.A.Ichnos 6, 141e149.

Crompton, R.H., Pataky, T.C., Savage, R., D’Août, K., Bennett, M.R., Day, M.H.,Bates, K.T., Morse, S., Sellers, W.I., 2011. Human-like external function of thefoot, and fully upright gait, confirmed in the 3.66 million year old Laetolihominin footprints by topographic statistics, experimental footprint-formationand computer simulation. J. R. Soc. Interface 9, 707e719.

D’Août, K., Meert, L., Van Gheluwe, B., De Clercq, D., Aerts, P., 2010. Experimentallygenerated footprints in sand: Analysis and consequences for the interpretationof fossil and forensic footprints. Am. J. Phys. Anthropol. 141, 515e525.

Darwin, C., 1871. The descent of man, and selection in relation to sex. J. Murray,London.

Day, M.H., Wickens, E.H., 1980. Laetoli Pliocene hominid footprints and bipedalism.Nature 286, 385e387.

Deino, A.L., 2011. 40Ar/39Ar dating of Laetoli, Tanzania. In: Harrison, T. (Ed.),Paleontology and Geology of Laetoli: Human Evolution in Context. Volume 1:Geology, Geochronology, Paleoecology and Paleoenvironment. Springer, NewYork, pp. 77e97.

DeMenocal, P.B., 2004. African climate change and faunal evolution during thePliocene-Pleistocene. Earth Planet. Sci. Lett. 220, 3e24.

Dentzien-Dias, P.C., Schultz, C.L., Bertoni-Machado, C., 2008. Taphonomy andpaleoecology inferences of vertebrate ichnofossils from Guar�a Formation (Up-per Jurassic), southern Brazil. J. S. Am. Earth Sci. 25, 196e202.

Dingwall, H.L., Hatala, K.G., Wunderlich, R.E., Richmond, B.G., 2013. Hominin stat-ure, body mass, and walking speed estimates based on 1.5 million-year-oldfossil footprints at Ileret, Kenya. J. Hum. Evol. 64, 556e568.

Duda, P., Zrzavy, J., 2013. Evolution of life history and behavior in Hominidae: To-wards phylogenetic reconstruction of the chimpanzeeehuman last commonancestor. J. Hum. Evol. 65, 424e446.

Ezquerra, R., Doublet, S., Costeur, L., Galton, P.M., P�erez-Lorente, F., 2007. Were non-avian theropod dinosaurs able to swim? Supportive evidence from an EarlyCretaceous trackway, Cameros Basin (La Rioja, Spain). Geology 35, 507.

Falkingham, P.L., 2012. Acquisition of high resolution three-dimensional modelsusing free, open-source, photogrammetric software. Palaeontol. Electron. 15,1e15.

Feibel, C.S., 2011. A geological history of the Turkana Basin. Evol. Anthropol. 20,206e216.

Gatesy, S.M., Middleton, K.M., Jenkins Jr., F.A., Shubin, N.H., 1999. Three-dimensionalpreservation of foot movements in Triassic theropod dinosaurs. Nature 399,141e144.

Gordon, A.D., Green, D.J., Richmond, B.G., 2008. Strong postcranial size dimorphismin Australopithecus afarensis: Results from two new resampling methods formultivariate data sets with missing data. Am. J. Phys. Anthropol. 135, 311e328.

Grabowski, M.W., Hatala, K.G., Jungers, W.L., Richmond, B.G., 2015. Body mass es-timates of hominin fossils and the evolution of human body size. J. Hum. Evol.85, 75e93.

Gurven, M., 2004. To give and to give not: The behavioral ecology of human foodtransfers. Behav. Brain Sci. 27, 543e583.

Haile-Selassie, Y., Melillo, S.M., Su, D.F., 2016. The Pliocene hominin diversityconundrum: Do more fossils mean less clarity? Proc. Natl. Acad. Sci. 113,6364e6371.

Hamilton, M.J., Milne, B.T., Walker, R.S., Burger, O., Brown, J.H., 2007. The complexstructure of hunter-gatherer social networks. Proc. R. Soc. B Biol. Sci. 274,2195e2202.

Hatala, K.G., Dingwall, H.L., Wunderlich, R.E., Richmond, B.G., 2013. The relationshipbetween plantar pressure and footprint shape. J. Hum. Evol. 65, 21e28.

K.G. Hatala et al. / Journal of Human Evolution 112 (2017) 93e104104

Hatala, K.G., Roach, N.T., Ostrofsky, K.R., Wunderlich, R.E., Dingwall, H.L.,Villmoare, B.A., Green, D.J., Braun, D.R., Richmond, B.G., 2016a. Footprints revealdirect evidence of group behavior and locomotion in Homo erectus. Sci. Rep. 6,28766.

Hatala, K.G., Wunderlich, R.E., Dingwall, H.L., Richmond, B.G., 2016b. Interpretinglocomotor biomechanics from the morphology of human footprints. J. Hum.Evol. 90, 38e48.

Hatala, K.G., Demes, B., Richmond, B.G., 2016c. Laetoli footprints reveal bipedal gaitbiomechanics different from those of modern humans and chimpanzees. Proc.R. Soc. B Biol. Sci. 283, 20160235.

Henderson, D., 2003. Footprints, trackways, and hip heights of bipedal dinosaurs etesting hip height predictions with computer models. Ichnos 10, 99e114.

Hill, K., 2002. Altruistic cooperation during foraging by the Ache, and the evolvedhuman predisposition to cooperate. Hum. Nat. 13, 105e128.

Hill, K., Barton, M., Hurtado, A.M., 2009. The emergence of human uniqueness:characters underlying behavioral modernity. Evol. Anthropol. 18, 187e200.

Hublin, J.-J., Ben-Ncer, A., Bailey, S.E., Freidline, S.E., Neubauer, S., Skinner, M.M.,Bergmann, I., Le Cabec, A., Benazzi, S., Harvati, K., Gunz, P., 2017. New fossilsfrom Jebel Irhoud, Morocco and the pan-African origin of Homo sapiens. Nature546, 289e292.

Kim, K.S., Kim, J.Y., Kim, S.H., Lee, C.Z., Lim, J.D., 2009. Preliminary report on hominidand other vertebrate footprints from the late Quaternary strata of Jeju Island,Korea. Ichnos 16, 1e11.

Kimbel, W.H., Delezene, L.K., 2009. “Lucy” redux: a review of research on Austral-opithecus afarensis. Yearb. Phys. Anthropol. 52, 2e48.

Kukihara, R., Lockley, M.G., 2012. Fossil footprints from the Dakota Group (Creta-ceous) John Martin Reservoir, Bent County, Colorado: new insights into thepaleoecology of the Dinosaur Freeway. Cretaceous Res. 33, 165e182.

Laporte, L.F., Behrensmeyer, A.K., 1980. Tracks and substrate reworking by terrestrialvertebrates in Quaternary sediments of Kenya. J. Sediment. Petrol. 50,1337e1346.

Leakey, M.D., 1981. Tracks and tools. Philos. Trans. R. Soc. 292, 95e102.Leakey, M.D., Hay, R.L., 1979. Pliocene footprints in the Laetolil Beds at Laetoli,

northern Tanzania. Nature 278, 317e323.Lee, R.B., 1968. What hunters do for a living, or, how to make out on scarce re-

sources. In: Lee, R.B., DeVore, I. (Eds.), Man the Hunter. Aldine PublishingCompany, Chicago, pp. 30e48.

Leutenegger, W., Shell, B., 1987. Variability and sexual dimorphism in canine size ofAustralopithecus and extant hominoids. J. Hum. Evol. 16, 359e367.

Lingham-Soliar, T., Broderick, T., Ait Kaci Ahmed, A., 2003. Closely associatedtheropod trackways from the Jurassic of Zimbabwe. Naturwissenschaften 90,572e576.

Liutkus-Pierce, C.M., Zimmer, B.W., Carmichael, S.K., McIntosh, W., Deino, A.,Hewitt, S.M., McGinnis, K.J., Hartney, T., Brett, J., Mana, S., Deocampo, D.,Richmond, B.G., Hatala, K., Harcourt-Smith, W., Pobiner, B., Metallo, A., Rossi, V.,2016. Radioisotopic age, formation, and preservation of Late Pleistocene humanfootprints at Engare Sero, Tanzania. Palaeogeogr. Palaeoclimatol. Palaeoecol.463, 68e82.

Lockley, M.G., 1994. Dinosaur ontogeny and population structure: interpretationsand speculations based on fossil footprints. In: Carpenter, K., Hirsch, K.F.,Horner, J.R. (Eds.), Dinosaur Eggs and Babies. Cambridge University Press, NewYork, pp. 347e370.

Lockley, M.G., Meyer, C.A., dos Santos, V.F., 1994. Trackway evidence for a herd ofjuvenile sauropods from the late Jurassic of Portugal. Gaia 10, 27e35.

Lockley, M.G., Houck, K., Yang, S.-Y., Matsukawa, M., Lim, S.-K., 2006. Dinosaur-dominated footprint assemblages from the Cretaceous Jindong Formation,Hallyo Haesang National Park area, Goseong County, South Korea: Evidence andimplications. Cretaceous Res. 27, 70e101.

Lockley, M.G., Mitchell, L., Odier, G.P., 2007. Small theropod track assemblages fromMiddle Jurassic eolianites of eastern Utah: paleoecological insights from duneichnofacies in a transgressive sequence. Ichnos 14, 131e142.

Lu, Z., Meldrum, D.J., Huang, Y., He, J., Sarmiento, E.E., 2011. The Jinnuishan homininpedal skeleton from the late Middle Pleistocene of China. HOMO J. Comp. Hum.Biol. 62, 389e401.

Marlowe, F., 2005. Hunter-gatherers and human evolution. Evol. Anthropol. 14,54e67.

Masao, F.T., Ichumbaki, E.B., Cherin, M., Barili, A., Boschian, G., Iurino, D.A.,Menconero, S., Moggi-Cecchi, J., Manzi, G., 2016. New footprints from Laetoli(Tanzania) provide evidence for marked body size variation in early hominins.eLife 5, 29.

Mastrolorenzo, G., Petrone, P., Pappalardo, L., Sheridan, M.F., 2006. The Avellino3780-yr-B.P. catastrophe as a worst-case scenario for a future eruption atVesuvius. Proc. Natl. Acad. Sci. 103, 4366e4370.

Matsukawa, M., Matsui, T., Lockley, M.G., 2001. Trackway evidence of herd structureamong ornithopod dinosaurs from the cretaceous Dakota group of North-eastern New Mexico, USA. Ichnos 8, 197e206.

McDougall, I., Brown, F.H., 2006. Precise 40Ar/39Ar geochronology for the upperKoobi Fora Formation, Turkana Basin, northern Kenya. J. Geol. Soc. Lond. 163,205e220.

Meldrum, D.J., 2004. Fossilized Hawaiian footprints compared with Laetoli hominidfootprints. In: Meldrum, D.J., Hilton, C.E. (Eds.), From Biped to Strider: TheEmergence of Modern Human Walking, Running, and Resource Transport.Kluwer Academic/Plenum Publishers, New York, pp. 63e83.

Mietto, P., Avanzini, M., Rolandi, G., 2003. Brief communications: Human footprintsin Pleistocene volcanic ash. Nature 422, 133.

Mitani, J.C., Watts, D.P., 2005. Correlates of territorial boundary patrol behaviour inwild chimpanzees. Anim. Behav. 70, 1079e1086.

Morse, S.A., Bennett, M.R., Liutkus-Pierce, C., Thackeray, F., McClymont, J., Savage, R.,Crompton, R.H., 2013. Holocene footprints in Namibia: the influence of sub-strate on footprint variability. Am. J. Phys. Anthropol. 151, 265e279.

Mountain, E.D., 1966. Footprints in calcareous sandstone at Nahoon Point. S. Afr. J.Sci. 62, 103e111.

Ostrom, J.H., 1972. Were some dinosaurs gregarious? Palaeogeogr. Palaeoclimatol.Palaeoecol. 11, 287e301.

Pablos, A., Pantoja-P�erez, A., Martínez, I., Lorenzo, C., Arsuaga, J.L., 2015. Metric andmorphological analysis of the foot in the Middle Pleistocene sample of Sima delos Huesos (Sierra de Atapuerca, Burgos, Spain). Quat. Int. 433, 103e113.

Plavcan, J.M., 2000. Inferring social behavior from sexual dimorphism in the fossilrecord. J. Hum. Evol. 39, 327e344.

Plavcan, J.M., Lockwood, C.A., Kimbel, W.H., Lague, M.R., Harmon, E.H., 2005. Sexualdimorphism in Australopithecus afarensis revisited: How strong is the case for ahuman-like pattern of dimorphism? J. Hum. Evol. 48, 313e320.

Pobiner, B.L., Rogers, M.J., Monahan, C.M., Harris, J.W.K., 2008. New evidence forhominin carcass processing strategies at 1.5 Ma, Koobi Fora, Kenya. J. Hum. Evol.55, 103e130.

Potts, R., 1998. Environmental hypotheses of hominid evolution. Yearb. Phys.Anthropol. 41, 93e136.

Potts, R., Faith, J.T., 2015. Alternating high and low climate variability: The context ofnatural selection and speciation in Plio-Pleistocene hominin evolution. J. Hum.Evol. 87, 5e20.

Raichlen, D.A., Gordon, A.D., Harcourt-Smith, W.E.H., Foster, A.D., Haas, W.R., 2010.Laetoli footprints preserve earliest direct evidence of human-like bipedalbiomechanics. PLoS One 5, e9769.

Rector, C.H., 1979. 5,000-year-old footprints on the Mojave River, California, USA.Antiquity 54, 149e150.

Reno, P.L., Lovejoy, C.O., 2015. From Lucy to Kadanuumuu: balanced analyses ofAustralopithecus afarensis assemblages confirm only moderate skeletal dimor-phism. PeerJ 3, e925.

Richmond, B.G., Harris, J.W.K., Mbua, E., Braun, D.R., Bamford, M., Bobe, R.,Green, D.J., Griffin, N.L., McCoy, J.T., Merritt, S., Pante, M., Pobiner, B., Carter-Menn, H., Chirchir, H., Kiura, P., Kibunjia, M., 2009. Divergence in hominin upperlimb anatomy in the early Pleistocene. Am. J. Phys. Anthropol. 138, 221.

Richmond, B.G., Hatala, K.G., Behrensmeyer, A.K., Bobe, R., Braun, D.R.,Dingwall, H.L., Green, D.J., Kiura, P., Villmoare, B.A., Wunderlich, R.E.,Harris, J.W.K., 2013. Hominin size, stature, and behavior based on 1.5-million-year-old footprints from Ileret, Kenya. PaleoAnthropology 2013, A32.

Roach, N.T., Hatala, K.G., Ostrofsky, K.R., Villmoare, B., Reeves, J.S., Du, A., Braun, D.R.,Harris, J.W.K., Behrensmeyer, A.K., Richmond, B.G., 2016. Pleistocene footprintsshow intensive use of lake margin habitats by Homo erectus groups. Sci. Rep. 6,26374.

Roach, N.T., Du, A., Hatala, K.G., Ostrofsky, K.R., Reeves, J.S., Braun, D.R.,Harris, J.W.K., Behrensmeyer, A.K., Richmond, B.G., 2017. Pleistocene animalcommunities at a 1.5 million-year-old lake margin grassland and their rela-tionship to Homo erectus paleoecology. J. Hum. Evol. submitted.

Roberts, D., Berger, L.R., 1997. Last interglacial (c. 117 kyr) human footprints fromSouth Africa. S. Afr. J. Sci. 93, 349e350.

Roberts, G., Gonzalez, S., Huddart, D., 1996. Intertidal Holocene footprints and theirarchaeological significance. Antiquity 70, 647e651.

Ruff, C.B., 1994. Morphological adaptation to climate in modern and fossil hominids.Yearb. Phys. Anthropol. 37, 65e107.

Ruff, C.B., 2010. Body size and body shape in early hominins - implications of theGona pelvis. J. Hum. Evol. 58, 166e178.

Ruff, C.B., Burgess, M.L., 2015. Howmuch more would KNM-WT 15000 have grown?J. Hum. Evol. 80, 74e82.

Smith, R.M.H., Marsicano, C.A., Wilson, J.A., 2009. Sedimentology and paleoecologyof a diverse Early Jurassic tetrapod tracksite in Lesotho, southern Africa. PAL-AIOS 24, 672e684.

Swedell, L., Plummer, T., 2012. A papionin multilevel society as a model for homininsocial evolution. Intl. J. Primatol. 33, 1165e1193.

Tuttle, R.H., Webb, D.M., Baksh, M., 1991. Laetoli toes and Australopithecus afarensis.Hum. Evol. 6, 193e200.

Watts, D., Muller, M., Amsler, S.J., Mbabazi, G., Mitani, J.C., 2006. Lethal intergroupaggression by chimpanzees in Kibale National Park, Uganda. Am. J. Primatol. 68,161e180.

Watts, D.P., Mitani, J.C., 2001. Boundary patrols and intergroup encounters in wildchimpanzees. Behaviour 138, 299e327.

Webb, S., Cupper, M.L., Robins, R., 2006. Pleistocene human footprints from theWillandra Lakes, southeastern Australia. J. Hum. Evol. 50, 405e413.

White, T.D., 1980. Evolutionary implications of Pliocene hominid footprints. Science208, 175e176.

Wilson, J.A., Carrano, M.T., 1999. Titanosaurs and the origin of “wide-gauge”trackways: a biomechanical and systematic perspective on sauropod locomo-tion. Paleobiology 25, 252e267.

Wrangham, R.W., Pilbeam, D., 2001. African apes as time machines. In:Galdikas, B.M.F., Briggs, N.E., Sheeran, L.K., Shapiro, G.L., Goodall, J. (Eds.), AllApes Great and Small. Kluwer Academic/Plenum Publishers, New York, pp. 5e17.